|
--- |
|
language: |
|
- be |
|
- cs |
|
- pl |
|
- ru |
|
- uk |
|
- zle |
|
- zlw |
|
tags: |
|
- translation |
|
- opus-mt-tc |
|
license: cc-by-4.0 |
|
model-index: |
|
- name: opus-mt-tc-big-zle-zlw |
|
results: |
|
- task: |
|
name: Translation rus-ces |
|
type: translation |
|
args: rus-ces |
|
dataset: |
|
name: flores101-devtest |
|
type: flores_101 |
|
args: rus ces devtest |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 23.1 |
|
- task: |
|
name: Translation ukr-ces |
|
type: translation |
|
args: ukr-ces |
|
dataset: |
|
name: flores101-devtest |
|
type: flores_101 |
|
args: ukr ces devtest |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 25.1 |
|
- task: |
|
name: Translation bel-pol |
|
type: translation |
|
args: bel-pol |
|
dataset: |
|
name: tatoeba-test-v2021-08-07 |
|
type: tatoeba_mt |
|
args: bel-pol |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 47.1 |
|
- task: |
|
name: Translation rus-ces |
|
type: translation |
|
args: rus-ces |
|
dataset: |
|
name: tatoeba-test-v2021-08-07 |
|
type: tatoeba_mt |
|
args: rus-ces |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 53.4 |
|
- task: |
|
name: Translation rus-pol |
|
type: translation |
|
args: rus-pol |
|
dataset: |
|
name: tatoeba-test-v2021-08-07 |
|
type: tatoeba_mt |
|
args: rus-pol |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 53.7 |
|
- task: |
|
name: Translation ukr-ces |
|
type: translation |
|
args: ukr-ces |
|
dataset: |
|
name: tatoeba-test-v2021-08-07 |
|
type: tatoeba_mt |
|
args: ukr-ces |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 58.0 |
|
- task: |
|
name: Translation ukr-pol |
|
type: translation |
|
args: ukr-pol |
|
dataset: |
|
name: tatoeba-test-v2021-08-07 |
|
type: tatoeba_mt |
|
args: ukr-pol |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 57.0 |
|
- task: |
|
name: Translation rus-ces |
|
type: translation |
|
args: rus-ces |
|
dataset: |
|
name: newstest2013 |
|
type: wmt-2013-news |
|
args: rus-ces |
|
metrics: |
|
- name: BLEU |
|
type: bleu |
|
value: 26.0 |
|
--- |
|
# opus-mt-tc-big-zle-zlw |
|
|
|
Neural machine translation model for translating from East Slavic languages (zle) to West Slavic languages (zlw). |
|
|
|
This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train). |
|
|
|
* Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.) |
|
|
|
``` |
|
@inproceedings{tiedemann-thottingal-2020-opus, |
|
title = "{OPUS}-{MT} {--} Building open translation services for the World", |
|
author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh}, |
|
booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation", |
|
month = nov, |
|
year = "2020", |
|
address = "Lisboa, Portugal", |
|
publisher = "European Association for Machine Translation", |
|
url = "https://aclanthology.org/2020.eamt-1.61", |
|
pages = "479--480", |
|
} |
|
|
|
@inproceedings{tiedemann-2020-tatoeba, |
|
title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}", |
|
author = {Tiedemann, J{\"o}rg}, |
|
booktitle = "Proceedings of the Fifth Conference on Machine Translation", |
|
month = nov, |
|
year = "2020", |
|
address = "Online", |
|
publisher = "Association for Computational Linguistics", |
|
url = "https://aclanthology.org/2020.wmt-1.139", |
|
pages = "1174--1182", |
|
} |
|
``` |
|
|
|
## Model info |
|
|
|
* Release: 2022-03-23 |
|
* source language(s): bel rus ukr |
|
* target language(s): ces pol |
|
* valid target language labels: >>ces<< >>pol<< |
|
* model: transformer-big |
|
* data: opusTCv20210807+bt ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge)) |
|
* tokenization: SentencePiece (spm32k,spm32k) |
|
* original model: [opusTCv20210807+bt_transformer-big_2022-03-23.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/zle-zlw/opusTCv20210807+bt_transformer-big_2022-03-23.zip) |
|
* more information released models: [OPUS-MT zle-zlw README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/zle-zlw/README.md) |
|
* more information about the model: [MarianMT](https://huggingface.co/docs/transformers/model_doc/marian) |
|
|
|
This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>ces<<` |
|
|
|
## Usage |
|
|
|
A short example code: |
|
|
|
```python |
|
from transformers import MarianMTModel, MarianTokenizer |
|
|
|
src_text = [ |
|
">>pol<< Это метафора.", |
|
">>pol<< Что вы делали?" |
|
] |
|
|
|
model_name = "pytorch-models/opus-mt-tc-big-zle-zlw" |
|
tokenizer = MarianTokenizer.from_pretrained(model_name) |
|
model = MarianMTModel.from_pretrained(model_name) |
|
translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True)) |
|
|
|
for t in translated: |
|
print( tokenizer.decode(t, skip_special_tokens=True) ) |
|
|
|
# expected output: |
|
# To metafora. |
|
# Co robiliście? |
|
``` |
|
|
|
You can also use OPUS-MT models with the transformers pipelines, for example: |
|
|
|
```python |
|
from transformers import pipeline |
|
pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-zle-zlw") |
|
print(pipe(">>pol<< Это метафора.")) |
|
|
|
# expected output: To metafora. |
|
``` |
|
|
|
## Benchmarks |
|
|
|
* test set translations: [opusTCv20210807+bt_transformer-big_2022-03-23.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/zle-zlw/opusTCv20210807+bt_transformer-big_2022-03-23.test.txt) |
|
* test set scores: [opusTCv20210807+bt_transformer-big_2022-03-23.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/zle-zlw/opusTCv20210807+bt_transformer-big_2022-03-23.eval.txt) |
|
* benchmark results: [benchmark_results.txt](benchmark_results.txt) |
|
* benchmark output: [benchmark_translations.zip](benchmark_translations.zip) |
|
|
|
| langpair | testset | chr-F | BLEU | #sent | #words | |
|
|----------|---------|-------|-------|-------|--------| |
|
| bel-pol | tatoeba-test-v2021-08-07 | 0.65517 | 47.1 | 287 | 1706 | |
|
| rus-ces | tatoeba-test-v2021-08-07 | 0.69695 | 53.4 | 2934 | 16831 | |
|
| rus-pol | tatoeba-test-v2021-08-07 | 0.72176 | 53.7 | 3543 | 21505 | |
|
| ukr-ces | tatoeba-test-v2021-08-07 | 0.73149 | 58.0 | 1787 | 8550 | |
|
| ukr-pol | tatoeba-test-v2021-08-07 | 0.74649 | 57.0 | 2519 | 13201 | |
|
| bel-ces | flores101-devtest | 0.41248 | 11.1 | 1012 | 22101 | |
|
| bel-pol | flores101-devtest | 0.42240 | 10.2 | 1012 | 22520 | |
|
| rus-ces | flores101-devtest | 0.50971 | 23.1 | 1012 | 22101 | |
|
| rus-pol | flores101-devtest | 0.48672 | 18.4 | 1012 | 22520 | |
|
| ukr-ces | flores101-devtest | 0.52482 | 25.1 | 1012 | 22101 | |
|
| ukr-pol | flores101-devtest | 0.48790 | 18.8 | 1012 | 22520 | |
|
| rus-ces | newstest2012 | 0.45834 | 18.8 | 3003 | 65456 | |
|
| rus-ces | newstest2013 | 0.52364 | 26.0 | 3000 | 57250 | |
|
|
|
## Acknowledgements |
|
|
|
The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland. |
|
|
|
## Model conversion info |
|
|
|
* transformers version: 4.16.2 |
|
* OPUS-MT git hash: 1bdabf7 |
|
* port time: Thu Mar 24 00:50:29 EET 2022 |
|
* port machine: LM0-400-22516.local |
|
|